prosím do čestiny
To evaluate our group decision support system methodology we have chosen the
recommendation domain because it is a clear example of group decision support
systems. Moreover, movie recommendation is a very accessible area with datasets
available and, more importantly, well known to users. The main hypotheses to
validate are:
H1. The multi-agent architecture of deliberative agents connected according to a
real social network improves the accuracy of standard “fully connected”
group recommenders.
H2. The personality and trust factors improve the performance of dis- tributed
group recommenders.
H3. Individual satisfaction increases when using this new group recom- mendation
technique because agents are able to argue about the item chosen.